dsmiley commented on PR #4320: URL: https://github.com/apache/solr/pull/4320#issuecomment-4460045934
RE: Removing `SolrClientCache` ... I'm unsure of; maybe. Even if we agree with that, that's too much for the 10.x cycle honestly, and furthermore wouldn't want to hold up this nice PR with such a move. It's not internal; anyone using solrj-streaming (e.g. at a client / java app; I don't mean from within Solr) is touching it and will be impacted by API changes. I wear that hat as a client of this at work. To address the security concern, minimally this PR must differentiate an existing user of that, and Solr itself, probably by using a different constructor. Backwards compatibility can be preserved... maybe not _total_ compatibility but at least for a common/typical construction of the SCC. I proposed a predicate and a `put` but we could even have something simpler -- a constructor that takes a Map that is the entirety of the available CloudSolrClients, which in practice would be of size 1 and this constructor would always be used from within Solr. > us creating a seperate "blocker" issue for 10.1 that is "Determine CLI commands that operate only on ZK with no Solr node running", and then carve off that thinking? I don't see that as a blocker. But I would like to see the CLI scope creep in this PR fully decoupled (removed) from this PR; it's too much and is a distraction to streaming expression oriented stuff (which is the basis of the rest of the PR even if it may not appear that way on the surface). -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
